多因子贡献率权重的城市精细人口空间化方法:以北京市为例

Spatial method of urban fine population with multifactor contribution rate weight:case of Beijing

以北京市作为研究区,以土地覆盖、河流道路、DEM、珞珈一号夜间灯光以及兴趣点各数据作为影响因子建立人口指标体系,针对人口空间化中多源数据权重确定方法复杂且主观性强的问题,提出以主成分分析法挖掘各项数据特征,以因子对指标评价体系的贡献率作为权重进行多因子加权融合,实现北京市30 m×30 m精细人口空间化.与Worldpop数据进行精度对比分析,结果表明:1)人口精细空间化结果的精度(R^(2)=0.85、E_(MR)为0.28)优于Worldpop数据的精度(R^(2)=0.67,E_(MR)为0.48);2)北京市人口主要分布在中心主城区内,且向外呈现波动性递减趋势,在外围出现多个郊区次中心,与兴趣点密度具有相似的空间分布;3)各类兴趣点因子被赋予了较高的权重,可作为北京人口空间分布的重要指示性因子.该研究方法可为其他城市精细人口空间化研究提供参考.

The study of urban population spatialization in fine scale can help to analyze characteristics and influencing factors of population distribution,and provide data support for regional sustainable development.A population index system was established for Beijing with land cover data,river and road data,DEM data,night light data of Luojia-01 and point of interest data as influence factors.To solve the problem of complex and subjective method of determining weight of multi-source data in population spatialization,an objective and scientific principal component analysis method was proposed to mine data characteristics.Contribution rate of factors to the index evaluation system was used as weight to carry out multifactor weighted fusion,to realize fine spatialization of 30 m×30 m population in Beijing.The accuracy of population spatialization(R^(2)=0.85,E_(MR)=0.28)was found to be better than Worldpop data(R^(2)=0.67,E_(MR)=0.48).The population in Beijing was found mainly distributed in the main urban area in the center,showing a decreasing trend of volatility outward,with multiple suburban sub centers in the periphery,a spatial distribution similar to density of interest points.All kinds of interest factors were given high weight,to be used as important indicator of spatial distribution of the Beijing population.This method might provide some reference for the study of spatial population in other cities.